mlsanityData to 7 Oct 2026

Millions of Co-purchases and Reviews Reveal the Spread of Polarization and Lifestyle Politics across Online Markets

arXiv 2201.06556 ↗First posted 17 Jan 2022cs.SI cs.AI cs.CLcited by 1

Not on the map. Why?

Abstract (arXiv)

Polarization in America has reached a high point as markets are also becoming polarized. Existing research, however, focuses on specific market segments and products and has not evaluated this trend's full breadth. If such fault lines do spread into other segments that are not explicitly political, it would indicate the presence of lifestyle politics -- when ideas and behaviors not inherently political become politically aligned through their connections with explicitly political things. We study the pervasiveness of polarization and lifestyle politics over different product segments in a diverse market and test the extent to which consumer- and platform-level network effects and morality may explain lifestyle politics. Specifically, using graph and language data from Amazon (82.5M reviews of 9.5M products and product and category metadata from 1996-2014), we sample 234.6 million relations among 21.8 million market entities to find product categories that are most politically relevant, aligned, and polarized. We then extract moral values present in reviews' text and use these data and other reviewer-, product-, and category-level data to test whether individual- and platform- level network factors explain lifestyle politics better than products' implicit morality. We find pervasive lifestyle politics. Cultural products are 4 times more polarized than any other segment, products' political attributes have up to 3.7 times larger associations with lifestyle politics than author-level covariates, and morality has statistically significant but relatively small correlations with lifestyle politics. Examining lifestyle politics in these contexts helps us better understand the extent and root of partisan differences, why Americans may be so polarized, and how this polarization affects market systems.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 10 of its 32 references that are in the map. First posted on arXiv 17 Jan 2022.

Builds on

10 of 32 references.

References · 10

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  3. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,282
  4. 2013Distributed Representations of Words and Phrases and their Compositionalitycited by 3,462
  5. 2018Deep contextualized word representationscited by 2,285
  6. 2019A Survey on Bias and Fairness in Machine Learningcited by 1,911
  7. 2016Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddingscited by 1,304
  8. 2016Ups and Downs: Modeling the Visual Evolution of Fashion Trends with One-Class Collaborative Filteringcited by 517
  9. 2019Lipstick on a Pig: Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Themcited by 268
  10. 2019Analyzing Polarization in Social Media: Method and Application to Tweets on 21 Mass Shootingscited by 67

Built on by

Cited by year

Citing papers per year: 2022: 1.

1 paper.